Metadata Education Project

Metadata education suggestions and materials for:

Learning Material | Preparatory topics | Complementary topics | Vocabulary

• What is metadata?
• Why is metadata important?
• Overview of the metadata content standard

• Data sources: determining fitness-for-use

Vocabulary definitions

• metadata document
• metadata element
• Identification Information section
• Data Quality Information
• Spatial Reference Information
• Spatial Organization Information
• Entity and Attribute Information
• Distribution Information

• Metadata documents: Distinguishing between quality metadata documents and incomplete documents
• Metadata elements: Distinguishing between quality metadata elements and incomplete elements

"Completeness" may be characterized as having descriptions for all elements of the dataset: identification information, data quality (including lineage), spatial organization, spatial reference, distribution information (if data is to be distributed), and metadata information. In addition, completeness means that each of these sections is filled out to a useful (not just a minimal) level of detail. For instance, here is an incomplete example of a Data Quality section. Note that though information has been filled in for each of the elements, the information is only minimal. This is another example of an incomplete section. Compare these to the good examples of Data Quality elements below.

Many elements of metadata are necessary only when appropriate for a particular dataset. For instance, if the data does not have a vertical component (such as height or depth), then it is not necessary to have an element for vertical horizontal positional accuracy.

In cases where a metadata document is missing one or more sections (especially Data Quality) or the elements under the sections contain only minimal information, then the metadata has been filled out primarily for the purposes of cataloging and not for determining fitness-for-use. In such cases, use the contact information provided either in the Identification Information or Distribution Information sections to call or otherwise contact the datasets' producers and/or distributors in order to get additional information required for determining fitness-for-use. If a producer/distributor cannot give you details about the dataset's attribute and positional accuracy, sources, process steps, structure and spatial reference, then there is simply not enough known about the dataset for it to be used with any sense of reliability.

Sources for example metadata were all taken from a survey of metadata documents drawn from the National Spatial Data Infrastructure Geospatial Clearinghouse in January 2000. If you are aware of additional examples, which you feel would be useful for education purposes, please contact us. Acknowledgments to the following metadata distributors for their examples:

• Cornell University Geospatial Information Repository
• Georgia GIS Data Clearinghouse
• Louisiana Geographic Information Center
• Maryland Geospatial Data Clearinghouse (UMBC)
• Wyoming Natural Resources Data Clearinghouse
• Utah Automated Geographic Reference Center
• Vermont Geographic Information Center

• Citation
  Only the title of dataset is returned from a search, in the initial listing. Therefore titles should be very descriptive, avoiding the use of codes, abbreviations and acronyms unless they are universally known. Some recommend that you go so far as to include other citation within the title, such as date, location, and even publisher. Many titles also include the primary source-scale of the data (e.g. 1:100,000 or 1:24,000).
• inadequate example: the title is a filename, and provides no useable information about the data set. Note that the originator is the same as the publisher. Though this is not incorrect, usually the publisher is the same as the distributor (see Distribution Information section).
• better example: the title also contains a file name and a keyword "environmental", but contains no information about the area the data set covers.
• Abstract/Purpose
• good example: includes a brief description with sources, scale, procedures, dates, updates and archive information.
• basic example: this example is really too basic. More information would be helpful.
• detailed example: this dataset is actually part of a detailed study using GIS, so background information is given about the study and how the dataset fits within the study framework
• Keywords
• good example: socioeconomic dataset
• good example: wildlife model dataset
• Use Constraints
• good example: basic
• good example: detailed, from a project with national standards: includes appropriate/inappropriate uses
• good example: detailed, includes statements about altering data and citation of data
• Native data set environment
  This is usually the name and version of the software the data was created in (e.g. ArcView 3.0, Arc/Info 7.0.2). In addition the platform (e.g. Win 95, Win NT, Unix Solaris, Unix SGI) may be included and the specific filename and path of the dataset.

Data Quality Information

• Attribute Accuracy:
• good example: rare case where accuracy was determined by field-checking
• good example: an example of how updates to the attribute accuracy were incorporated (with dates) as the data was improved.
• special example: in modeled datasets, attribute accuracy sometime does not apply since the process is subjective or derivative.
• special example: reference is made to individual source data sets within the Lineage section. This is acceptable, but not ideal.
• Positional Accuracy:
• incomplete example: data recorded by a GPS should include not just final accuracy but how measurements were derived
• good example: accuracy of data recorded by a GPS along with how measurements were derived
• basic example: accuracy of data digitized from hardcopy maps
• good example: slightly more detailed accuracy of data digitized from hardcopy maps
• good example: estimated accuracy of data created from a legal description conversion process
• special example: a derived data set, combined in a spatial analysis procedure from multiple data sources, each with varying horizontal accuracy.
• special example: in rare cases, horizontal positional accuracy does not apply
• special example: reference is made to individual source datasets within the Lineage section. This is acceptable, but not ideal.
• Logical consistency:
• good example
• Completeness report
• good example
• adequate example
• Source information
• good example: includes detailed citation, source-scale denominator and source contribution. Only field missing is type of source media (e.g. digital, paper, magnetic)
• Process Steps
• good example: good logical organization and complete reference information (commonly missing for process steps)
• good example: carefully detailed steps
• inadequate example: no information on how process steps were performed
• inadequate example: no information on how process steps were performed

• good example: a raster structure
• good example: a vector structure with polygon topology
• incomplete example: a vector structure but no details of structure or topology provided

• Good examples: includes resolution, parameters, and datum:
• Universal Transverse Mercator coordinate system (raster data)
• Lambert Conformal Conic (vector data)
• Geographic coordinate system (lat/long in decimal degrees)
• State Plane coordinate system
• Incomplete example: unknown resolution; missing datum information

• inadequate example: brief explanation of attribute field names, but no information about their content
• basic example: brief explanation of attribute field names, and their codes, with reference to a database file that contains more detailed information
• good example: entity/attribute information in overview format, which is probably easiest to read for those unfamiliar with metadata formatting
• good example: entity/attribute information in detailed format, more difficult to read for those unfamiliar with metadata formatting, but well-organized in a formal format, beginning with the entity and following up with the entity's attributes

• Liability
• good example
• detailed example
• Standard Order Process
• good example: several different available methods/formats

Back to Course Topics